`Identification of Plausible Folded Conformers
`
`GILDA LOEW, JACK COLLINS, PHILIP PAYNE, AMRIT K. JUDD, and
`KEVIN H. WACKNOW
`SRI International, 333 Ravenswuod Avenue, Menlo Park, Calijurnia 94025, USA
`
`Abstract
`
`@-Endorphins are 31 amino acid endogenous opioid peptides with high receptor affinity and antinocicep-
`tive acvitity. Because of their importance as neurohormones and the significant experimental effort that has
`been made to understand their structure activity profiles, we have begun to develop procedures that could
`be useful ftrst to identify low-energy conformers of P-endorphins and ultimately their bioactive form. In
`the initial studies reported here, we have identified plausible initial structures of the full peptide by calcu-
`lating and comparing the conformational preference of all possible extended tetrapeptide fragments of
`P-endorphin starting from each of the first 28 residues. Comparisons of fragment energies suggested two
`types of compact folded @-endorphin conformers were plausible: a helix-turn-helix and an antiparallel
`@-sheet conformer. These structures, as well as an extended a-helical and @-strand conformer, were assem-
`bled and total geometry optimization performed using the empirical-energy-based program AMBER. The
`results yield an a-helical structure as the lowest energy form consistent with recently reported NMR stud-
`ies of @-endorphin. The two more compact folded structures obtained, however, are reasonable starting
`conformations for further planned molecular dynamics simulation studies and could yield competing low-
`energy structures as candidates for the bioactive form of these peptides.
`
`Introduction
`@-Endorphins, 31 amino acid fragments of a larger prohormone, are potent en-
`dogenous opioid peptides with high receptor affinity and antinociceptive activity [ 11.
`That the sequence of @-endorphins is remarkably conserved across a variety of spe-
`cies is an indication that more than just the amino terminal met-enkaphalin-like por-
`tion of the peptides is important for activity. Since its discovery, about one hundred
`different analogs of @-endorphins have been synthesized in an attempt to determine
`the importance of individual residues and regions to the affinity and activity of the
`peptide [ 1-31. These extensive structure-activity studies include replacement, omis-
`sion, and addition of residues and incorporation of disulfide bridges [l-31. In another
`approach, variations in residues in the 13-31 regions were made based on the hy-
`pothesis [4] that all that is required for opioid activity in @-endorphins is that this re-
`gion form an amphiphilic helical structure. While the resulting analogs have shown
`binding affinity ranging from less than micromolar to nanomolar, and varying effica-
`cies as analgesics, the fundamental stereoelectronic properties that determine these
`variations have not yet been identified.
`
`International Journal of Quantum Chemistry, Quantum Biology Symposium, 15, 055-066 (1988)
`CCC 0360-8832/88/010055-12$04.00
`0 1988 by John Wiley & Sons, Inc.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 1 of 12
`
`
`
`56
`
`LOEW ET AL.
`
`This lack is not surprising since the P-endorphins belong to a class of intermediate
`size bioactive peptides for which characterization of conformation profiles is most
`difficult. Such peptides pose discouraging difficulties for each of the three new disci-
`plines: x-ray crystal structure determination, NMR studies, and theoretical energy-
`conformational studies, that in principle could be most useful for such studies. They
`are, in general, difficult to crystallize, have conformational flexibilities at room tem-
`perature, and have many possible stable structures, i.e., many local conformational
`minima of varying relative energies.
`Recently, an NMR study of P-endorphin has been published [ 5 ] . While confonna-
`tional flexibility in water did not allow analysis of the spectra, dilution with methanol
`did. The extensive analysis made of the NMR results, in water-methanol solution, in-
`cluding NOE spectra, were consistent with a predominantly a-helical structure. The
`NMR data do not preclude, however, a possible turn near residues 10-15. However,
`these studies are only a first step in addressing the question of the bioactive conform-
`ers of P-endorphins, i.e., the form in which they bind to opioid receptors. There are
`undoubtedly a number of lower-energy candidate structures and the environment at
`the opioid receptor binding site is most likely devoid of bulk solvent.
`The two more formidable obstacles to energy conformation studies of peptides of
`the size of @-endorphin, are the existence of large numbers of stable conformers, i .e.,
`the multiminimum problem, and the practical difficulties of constructing conformers
`which are good initial approximations to these minima. Human P-endorphin, for ex-
`ample, has 184 nonhydrogen torsion angles which render impractical the search
`strategies routinely used for smaller peptides such as nested rotations and “buildup”
`from low energy conformers of single amino acids. Thus entirely different procedures
`must be developed to search confonnational space of these peptides for low-energy
`conformers.
`The most common approach used to relate amino acid sequence to secondary struc-
`ture in large peptides is based on statistical analysis of x-ray structure data of
`proteins [6]. By contrast, in the work reported here, a novel search strategy, based on
`comparisons of calculated optimized energies of sequential peptide fragments was de-
`veloped to help identify plausible secondary structure regions for P-endorphin. These
`fragments were then used as guides to fold the peptide into a small number of qualita-
`tively different conformations, i.e,, into a set of tertiary structures which were then
`subjected to complete geometry optimizations.
`The results obtained thus far indicate that the search strategy developed could be
`useful to construct initial conformers of bioactive peptides in general and to address
`aspects of the protein folding problem.
`
`Methods and Procedures
`As a guide to construction of plausible conformers of P-endorphin, 28 overlapping
`elongated and derivatized tetrapeptide fragments were constructed of the form:
`CH,CONH-ala-(res,-resi,,)ala-CONHCH3, i = 1-28. Each of the 28 fragments
`were constructed in 6 idealized backbone conformations corresponding to an a-helix,
`a &strand, and four p-turns; I, 1’, 11, and 11’. The tetramers were extended to hexam-
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 2 of 12
`
`
`
`P-ENDORPHINS
`
`57
`
`ers by adding an alanine on each side to allow even-handed comparisons of the ener-
`gies of these various secondary structures since a minimum of 6-, 5 - , and
`4-contiguous residues are required for favorable H-bonding in a-helical, P-strand and
`@-turn fragments respectively. The N-terminal and C-terminal ends were appropri-
`ately derivatized by N-acetyl groups and carboxy N-methyl to more realistically
`mimic the conformational behavior of the segment as part of the larger peptide chain.
`Facile construction of these fragments was possible using the capabilities of an inter-
`active structure generating program called MOLECULE, described elsewhere (71,
`which has a library of single amino acid structures and the ability to automatically
`generate peptides of a chosen backbone conformation with extended side chain tor-
`sion angles. For fragments 10, 11, 12, and 13, which include the proline residue 13,
`two optimized proline ring geometries called PROu and PROd were used. In frag-
`ment 11, in which Pro13 is the second residue in the p-turn, only turn types I and 11’
`are possible. For fragment 12, in which Pro13 is the first residue in the turn, only
`turn types 11 and I1 are possible. All of the initial structures generated for the
`28 fragments were optimized in two steps, side chain angles only and then full tor-
`sion angle optimization using a quasi-Newton-Raphson energy minimization pro-
`gram called PEP that was developed in our laboratory. The five-term empirical
`energy expression in the program called ECCEP [8] formed the basis of this opti-
`mization. It contains contributions from electrostatic, hydrogen-bonded, dispersion,
`repulsion, and torsion angle potentials, and is described in detail elsewhere [8,9].
`In a buildup procedure similar to that used by Scheraga and co-workers [ 101, plau-
`sible folded structures of @endorphin were constructed by linking energy optimized
`fragments corresponding to different types of secondary structures. This process was
`not automatic, but involved extensive use of graphics capabilities, and distance opti-
`mization to obtain interfragment side-chain conformers which eliminated major steric
`repulsions.
`A set of 5 initial conformers generated for P-endorphin were energy optimized us-
`ing the empirical energy expression contained in the program AMBER [ 111 described
`in detail elsewhere. This program allows total geometry optimization and contains a
`7-term energy expression including bond angle and bond length variations in addition
`to the torsion angle variation and four other types of terms similar to those in the
`ECCEP potential. In these calculations, all atoms were explicitly included; with a
`nonbonding atom distance cutoff of 10 A. A distance-dependent dielectric, E = Y,
`was used. In this program, the polar side chain residues are assumed to be ionized.
`To achieve charge neutrality, salt bridges were formed between oppositely charged
`nearby amino acid side chains and the remainder of the charged residues (Lys) were
`neutralized by addition of negative counterions with Van Der Waal’s radii of 3 A.
`This procedure resulted in salt bridge formation in the antiparallel @-sheet between
`Lys29-Glu31 and Lys19-Glu8 and counterions at Lys9, 24, and 28. In the P-strand,
`salt bridges were formed between Glu8-Lys9; Lys29-Glu3 1, and counterions were
`placed at Lysl9, 24, and 28. In the a-helical structure, salt bridges were formed be-
`tween Glu8-Ly9, Lys24-Glu31, and counterions placed at Lysl9, 28, and 25. Fi-
`nally, in the turn-helix-turn-helix-turn
`structure in the left-handed helix, salt
`bridges were formed between Gly8-Ly9, Lys28-Glu3 1, and counterions were placed
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 3 of 12
`
`
`
`58
`
`LOEW ET AL.
`
`at Lysl9, 24, and 29; while in the right-handed helix salt bridges were formed be-
`tween Glu8-Lys9, Lys29-Glu31, and counterions placed at Lysl9, 24, and 28.
`
`Results and Discussion
`The optimized energies of the 28 extended tetrameric fragments of @-endorphin in
`the different backbone conformations are summarized in Table I relative to that of the
`alpha helical form. As shown in this table, a /?-turn conformation is preferred for the
`enkephalin portion of the peptide. These results are consistent with our own [ 121 and
`other previously reported energy-conformation studies of both tetra and pentapeptides
`and NMR studies of metenkephalin [ 13, 141 in which evidence for both a gly-gly and
`a gly-phe @-bends have been reported. For fragments 10, 11, 12, and 13, which in-
`clude the proline 13 residue, one proline ring geometry, Pro-D, definitely favored an
`optimized alpha helical structure with some deviations from ideal backbone angles.
`No turns were possible with this proline ring geometry. For the other proline ring
`geometry however, PI-type turns, again deviating from ideal, were favored for
`fragments 10, 11, and 12, identifying a crucial turn region in the middle of the
`@-endorphin sequence which could allow a highly folded structure. A third turn re-
`gion involving C-terminal residues 28-3 1 was also suggested by these results, though
`an a-helix is somewhat favored.
`In addition to identification of possible turn regions, the results suggest that the re-
`maining contiguous region of the peptide, i.e., residues 5-10, and residues 15-27 are
`in modified alpha helical rather than @-strand conformations. Thus the most plausible
`folded conformer of @-endorphins, is predicted to be of the helix-turn-helix pattern
`with possible additional turns at both the N-terminal and C-terminal end and an inter-
`nal turn beginning at residue 11 or 12. Two highly folded structures of this type were
`constructed, one with a right-handed and the other a left-handed helix, to explore the
`effect of the sense of the helical portions of the conformation and energy. A totally
`helical structure was also constructed as a possible variation of these compact folded
`structures.
`While @-strands, with only a few exceptions, were not energetically favorable sec-
`ondary structures for fragments, nevertheless, the possibility existed that a fully ex-
`tended @-strand could be a low-energy conformer of the full peptide since P-strands
`are likely to have less hindered interfragment interactions than a-helices. It is also
`possible that a highly folded @sheet structure, which can be formed from antiparallel
`@-strands by interstrand H-bonding could be energetically favorable if the energy
`gained from interstrand interactions outweighs the favorable a-helical versus P-strand
`contiguous domain energies. These possibilities were explored by including among
`the initial conformers to be optimized an extended @-strand and an antiparallel
`P-sheet conformer with turns at residues 1-4, 11-14, and 23-26.
`The results of total geometry optimization of the five types of conformers of
`@-endorphins using the AMBER program are summarized in Table 11. These opti-
`mized conformations are shown in Figure 1; their backbone conformations in Fig-
`ure 2, and the corresponding backbone torsion angles are listed in Tables 111-VI. As
`seen in Tables I11 and IV, fairly regular a-helical [Fig. l(a)], and @strand [Fig. ltb)]
`structures are retained after optimization. This is not true for the more folded struc-
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 4 of 12
`
`
`
`&ENDORPHINS
`
`59
`
`TABLE I. Optimized energies" of &-endor-
`phin fragments CHJCONH-ala[Res,-
`Res,,,,]-ala-CONHCH,.
`
`i
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`
`11
`
`12
`
`13
`
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`Fragment
`
`A&,
`
`A&
`
`S r
`GlY
`GlY
`Phe
`Met
`Thr
`Ser
`Glu
`LYS
`SerU'
`SerD
`GlnU'.'
`GlnD
`ThrV~"
`ThrD
`PrOU'
`ProD
`Leu
`Val
`T h
`Leu
`Phe
`LYS
`Asn
`Ala
`Ile
`Ile
`LY s
`Asn
`Ala
`TYr
`Lys-8
`
`3
`7
`21
`20
`17
`14
`16
`17
`23
`- 1
`2
`-6
`1
`0
`2
`7
`7
`21
`14
`12
`15
`14
`16
`15
`16
`16
`14
`14
`14
`15
`15
`3
`
`aAE in kcal/mol relative to energy of a-
`helix.
`b( ) = optimized turn with lowest energy
`of I, 1', 11, 11'.
`'Two proline ring geometries called U
`and D were used in these fragments.
`dOnly p-turn types I and 11' possible with
`Prol3.
`'Only P-turn types I and I1 possible with
`Prol3.
`'Turns sterically hindered.
`8The C-terminal fragment is Lys-Lys-
`Gly-Gln.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 5 of 12
`
`
`
`60
`
`LOEW ET AL
`
`TABLE 11. Optimized energies for four types p,-endorphin structures.
`
`Helix-turn- helix
`
`Energy"
`
`aR -helix
`
`aR
`
`AE total
`bond
`nonbond
`1-4 nb
`angle
`eel
`1-4 eel
`dihed
`hbond
`
`0.0
`0.0
`0.0
`0.0
`0.0
`0.0
`0.0
`0.0
`0.0
`
`41.7
`0.4
`- 16.8
`1.2
`9.4
`4.9
`24.9
`17.1
`0.4
`
`"All energies relative to a-helix.
`
`f f L
`
`125.5
`2.0
`- 1.0
`- 5.5
`35.7
`16.9
`27.9
`53.5
`- 3.9
`
`P-sheet
`
`P-strand
`
`89.3
`0.4
`4.6
`- 2.7
`5.4
`59.8
`0.5
`20.8
`0.4
`
`247.4
`- 1.3
`62.4
`5.1
`4.1
`65.1
`11.9
`- 0.7
`0.7
`
`Figure 1. Five optimized conformers of P-endorphin showing both backbone and side
`chains: (a) P-strand, (b) a-helix, (c) antiparallel P-sheet, (d) PT-ff~-Pr-aR-Pr, (d) PT-aL-
`PT-~L-PT.
`
`tures [Fig. l(c)-(e)] in which extensive interstrand side chain interaction causes large
`deviations from ideal secondary structure. As seen in Table V [Fig. l(c)], in the
`antiparallel P-sheet structure residues 1-4 form a distorted PII' turn; 11-14, a dis-
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 6 of 12
`
`
`
`&ENDORPHINS
`
`61
`
`Figure 2. Five optimized backbone conformers of P-endorphin: (a) P-strand, (b) a-helix,
`(c) antiparallel @sheet, (d) PT-aR-PT-aR-&, (d) Pr-aL-/3-aL-PT.
`
`torted PI11 turn, and residues 23-26 a distorted PI turn. The remainder of the struc-
`ture consists of antiparallel P-strand distorted from ideal values to minimize steric
`hindrance. Similar distortions from ideal secondary structure can be seen in the other
`highly folded structures.
`As seen in Table V, the left-handed helical structure, /3T-aL-j3T-aL-PT, has a pH’-
`type turn in residues 1-4 and 11-14, with the remainder of the structure in distorted
`has a
`aL helical form. The right-handed helical compact structure ( P T - a R - P T - a R - P T )
`PII’ turn at the N-terminal and C-terminal segments, a very broad turn around Prol3,
`with the remaining segments distorted aR -helices.
`As seen in Table 11, the alpha helical structure is the most stable. Of the more
`folded structures, our results thus far favor the P r ~ R - P r ~ R - / 3 T structure. This
`structure is significantly stabilized over the left-handed helical folded structure and
`the antiparallel P-sheet. While the interstrand interactions in the P-sheet has con-
`siderably lower energy relative to the P-strand, the energy gained by interstrand
`H-bonding was not sufficient to make it the more favored compact structure.
`The finding that the two lowest energy conformers are predominantly a-helical is
`in agreement with the recently published NMR studies of P-endorphin which also in-
`fers large helical regions [5]. Other features of the structures which are consistent
`with those observed NMR spectra are a turn at the C-terminal end and a relatively
`close distance of the Phe 18 side chain with the delta methyl group of isoleucine 23
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 7 of 12
`
`
`
`62
`
`LOEW ET AL.
`
`TABLE 111. Backbone torsion angles for
`a-Helicala &-endorphin structure.
`
`4
`
`0
`- 49
`-54
`-64
`- 60
`- 67
`-56
`-55
`- 64
`- 73
`- 50
`- 56
`- 64
`- 54
`-61
`- 60
`- 59
`-55
`- 53
`- 59
`- 60
`-57
`- 54
`- 60
`-64
`- 62
`-51
`-59
`- 60
`-87
`
`h
`
`-42
`-43
`- 49
`- 49
`-47
`- 36
`-52
`-52
`-31
`- 65
`- 56
`- 62
`-43
`-48
`-45
`- 49
`-49
`- 50
`-51
`- 44
`- 47
`- 48
`-51
`- 40
`-41
`- 54
`- 36
`-27
`- 29
`-57
`
`0,
`
`176
`175
`179
`- 177
`178
`172
`178
`-180
`171
`-171
`-171
`175
`- 175
`177
`179
`179
`- I77
`176
`178
`- 179
`176
`173
`179
`179
`176
`173
`180
`174
`172
`-176
`
`Try1
`Gly2
`Gly3
`Phe4
`Met5
`Thr6
`Ser7
`Glu8
`Lys9
`SerlO
`Glal 1
`Thrl2
`Pro13
`Leu14
`Val 15
`Thrl6
`Lev17
`Phel8
`Lysl9
`Asn20
`Ah21
`11122
`I1123
`Lys24
`Am25
`Ala26
`Try22
`Lys28
`Lys29
`Gly30
`
`'Ideal torsion angle values for a-helix
`are +c = -72", JIt = -54". Fairly regular
`structure retained after optimization.
`
`(5 A) which is consistent with the upfield shift of this methyl group due to ring cur-
`rent perturbations. These results also support the growing experimental evidence that
`small peptides can fold with regular secondary structure patterns most often favoring
`those with predominantly a-helical regions [ 151. While it is premature to select these
`conformers as the bioactive form in which @endorphins bind to opioid receptors, the
`result thus far also supports the hypothesis that structures with alpha helical segments
`in the regions of residues 16-27 could be involved [4].
`In conclusion, the results obtained thus far indicate that P-endorphin can fold into
`conformers with clearly defined structural motifs that can be inferred from energy op-
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 8 of 12
`
`
`
`&ENDORPHINS
`
`63
`
`TABLE IV. Backbone torsion angles for op-
`timized @-strand" conformer fin -endorphin.
`
`Try 1
`Gly2
`GIy3
`Phe4
`Met5
`T h 6
`Ser7
`Glu8
`Lys9
`SerlO
`Glal 1
`Thrl2
`Pro13
`Leu14
`Val15
`Thrl6
`Lev17
`Phel8
`Lysl9
`Am20
`Ah21
`I1122
`I1123
`Lys24
`As1125
`Ala26
`Try22
`Lys28
`Lys29
`Gly30
`
`0
`- 173
`- 178
`- 164
`-157
`- 155
`-174
`- 157
`- 153
`- 152
`-169
`- 156
`- 77
`- 67
`-131
`- 153
`- 166
`- 152
`- 158
`- 147
`- 149
`- 160
`-151
`- 142
`- I29
`- 109
`- 169
`-163
`- 160
`- 170
`
`174
`173
`171
`I59
`154
`175
`156
`158
`139
`159
`155
`141
`65
`126
`154
`172
`153
`152
`143
`145
`148
`153
`150
`129
`126
`169
`152
`153
`158
`169
`
`177
`180
`178
`175
`168
`177
`172
`-170
`179
`-175
`176
`-176
`173
`172
`172
`-179
`173
`175
`171
`176
`177
`176
`-179
`175
`170
`176
`180
`179
`173
`-179
`
`'Ideal values for fi-strand:
`= -140",
`Jli = 135". Fairly regular structure main-
`tained except for a "kink" near Prol3.
`
`timized fragments and are not random conformations. While an a-helix appears to be
`favored thus far, it is possible that additional side chain variations and some changes
`in backbone structure will lead to lower energy forms of the more compact folded
`structures. We are continuing to explore the conformational behavior of this large
`peptide, by molecular dynamic studies using each of the optimized folded conformers
`obtained here as starting configurations. The results should lead to additional varia-
`tions and refinements of the lowest energy conformers among which the bioactive
`form will ultimately be identified.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 9 of 12
`
`
`
`64
`
`LOEW ET AL.
`
`TABLE V. Backbone torsion angles from op-
`timized antiparallel @-sheeta @. -endorphin
`conformer.
`
`Try1
`Gly2
`Gly3
`Phe4
`Met5
`Thl6
`Ser7
`Glu8
`Lys9
`Ser 10
`Glall
`Thrl2
`Pro13
`Leu14
`Val 15
`Thrl6
`Lev17
`Phel8
`Lysl9
`Am20
`Ah21
`11122
`I1123
`Lys24
`Am25
`Ala26
`Try22
`Lys28
`Lys29
`Gly30
`
`0
`70
`-47
`-134
`- 145
`- 134
`- 179
`- 75
`- 71
`- 49
`- 165
`- 44
`55
`-171
`54
`-124 - 147
`-131
`- 96
`- 58
`-138
`-126
`- 87
`- 50
`- 135
`-162
`- 122
`- 146
`- 57
`- 81
`
`36
`- 73
`- 54
`148
`164
`146
`53
`61
`86
`129
`96
`- 46
`-39
`107
`135
`157
`114
`75
`67
`145
`110
`55
`149
`- 28
`40
`107
`109
`126
`98
`45
`
`179
`177
`156
`174
`173
`179
`172
`178
`158
`- 165
`- 165
`176
`- 179
`172
`173
`- I61
`166
`179
`-168
`172
`168
`-179
`-175
`173
`168
`-170
`162
`156
`-155
`- 177
`
`“Residue 1-4 form a distorted @II’ turn;
`11-14 a distorted @IU turn and 23-26 a dis-
`torted PI turn. The remainder of the struc-
`tures are antiparallel @-strands distorted
`from ideal values to minimize steric hin-
`drance of side chains.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 10 of 12
`
`
`
`P-ENDORPHINS
`
`65
`
`TABLE VI. Backbone torsion angles for optimized Pr-a-Pr-a-Pr conformers of Ph-
`endorphin.
`
`a,-helix
`
`A
`
`aL -helix
`
`B
`
`175.07
`14.55
`0.00
`-170.91
`166.84
`0.00
`Tyr 1
`- 72.17
`62.12 - 91.29
`Gly2
`177.48
`80.01
`172.59
`177.69 - 57.32 - 46.40
`170.44
`52.37
`-139.84
`Gly3
`137.07 - 175.35
`- 138.65 - 168.99 - 168.72 - 126.26
`Phe4
`- 39.92 - 48.49
`Met5
`-177.08
`66.72
`107.31 - 159.85
`94.56 - 16.61
`- 67.04 - 46.79
`146.36
`173.00
`Thr6
`- 57.56 - 47.78
`178.21
`-135.73
`121.72
`-140.54
`Ser7
`- 55.36 - 27.78
`78.87
`138.77
`-156.89
`Glu8
`169.51
`L Y S ~ - 80.80 - 58.67
`-172.73
`40.97
`41.57
`-165.67
`- 46.79 - 38.86
`178.05
`-170.12
`SerlO
`177.91
`151.34
`- 53.57 - 39.69
`63.74
`63.04
`173.64
`-160.56
`Glnll
`121.01 - 53.29
`- 51.19 - 51.65
`165.57
`-156.72
`Thrl2
`- 48.67 - 66.80
`- 48.18
`-175.21
`- 42.25
`166.03
`Pro13
`147.14
`166.78
`Leu14
`-101.47
`125.96
`-172.93
`-114.39
`- 45.49 - 49.41
`Vd15
`179.48
`-116.06
`130.43
`-138.00
`- 60.55 - 41.66
`-156.63
`69.16
`147.67
`Thrl6
`179.36
`Leu17 - 69.75 - 37.15
`51.94
`-159.97
`169.09
`53.11
`- 53.14 - 56.53
`99.69
`-173.12
`-176.53
`156.06
`Phel8
`84.27
`-145.84
`- 56.55 - 47.29
`177.01
`70.58
`Lysl9
`96.92
`-153.17
`Am20 - 61.43 - 52.82
`85.46
`-171.13
`104.92
`-147.21
`177.68
`74.52
`- 64.62 - 41.16
`Ala21
`- 63.07 - 37.08
`115.34
`-167.45
`62.85
`174.04
`Ile23
`87.58 - 166.72
`- 66.89 - 44.58
`172.39
`64.95
`Ile23
`Ly~24 - 51.91 - 53.07
`134.17
`-175.08
`81.09
`172.41
`60.47
`-147.68
`Asn25 - 49.34 - 44.32
`176.36
`59.48
`- 59.75 - 45.72
`273.47
`119.24
`92.84
`-136.60
`Ala26
`116.96 - 162.93
`-174.93
`61.47
`- 58.20 - 44.27
`Try27
`119.22
`-176.12
`Lys28 - 69.53 - 39.51
`-170.96
`74.52
`67.84 - 71.70
`169.79
`- 21.99
`172.64
`Ly~29
`-137.88
`- 166.89
`171.61 - 79.45
`- 63.34 - 43.50
`60.75
`Gly30
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 11 of 12
`
`
`
`66
`
`LOEW ET AL.
`
`Acknowledgment
`Support for this work from NIDA grant DA02622 is gratefully acknowledged. We
`are also grateful the use of the San Diego Supercomputer Center Class VI computers
`and the helpful guidance and support of the SDSC staff.
`
`Bibliography
`
`[I] C. H. Li, in Hormonal Proteins andPeptides, C. H. Li, Ed. (Academic, New York, 1980), p. 10.
`[2] P. Nicholas, R. G. Hammonds, Jr., and C. H. Li, Proc. Natl. Acad. Sci. U.S,A. 81, 3074 (1984).
`[3] P. Nicholas, R. G. Hammonds, Jr., and C. H. Li, Proc. Natl. Acad. Sci. U.S.A. 79, 2191 (1982).
`[4] J. W. Taylor and E.T. Kaiser, Pharmacol. Rev. 38, 291 (1986).
`[5] 0. Lichtarge, 0. Jardetsky, and C. H. Li, Biochemistry 26, 5916 (1987).
`[6] P. Y. Chou and G. D. Fasman, J. Mol. Biol. 115, 135 (1977).
`[7] J. T. Egan, J. Hart, S. K. Burt, and R. D. MacElroy, Comput. Graphics 6, 1977 (1982).
`[8] F. A. Momany, R. F. McGuire, A. W. Burgess, and H. A. Scheraga, J. Phys. Chem. 79, 7381
`(1975).
`[9] G. Nemethy, M. S. Pottle, and H. A. Scheraga, J. Phys. Chem. 87, 1883 (1983).
`[lo] M. Vasquez and H.A. Scheraga, Biopolymers 24, 1437 (1985).
`[ l l ] J. S. Weiner, P. A. Kollman, D. A. Case, U. C. Singh, C. Ghio, G . Alagona, S. Profeta, and P. J.
`Weiner, I. Am. Chem. SOC. 106, 765 (1984).
`[I21 G. Loew, H. Hashimoto, L. Williamson, S. Burt, and W. Anderson, Molec. Pharm. 22, 467 (1982).
`[I31 G. Gupta, FEBS Letters 198, 245 (1986).
`[I41 2. Li and H. A. Scheraga, Proc. Natl. Acad. Sci. U.S.A., 84, 6611 (1987).
`[I51 S. Marqusee and R. L. Baldwin, “The Protein Folding Problem,” Abstracts, AAAS Meeting, Boston
`MA, February 1988.
`
`Received May 3, 1988.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1076 Page 12 of 12